EGFR c.2303G>T, p.Ser768Ile

Reasoning

The observed EGFR NM_005228.5:c.2303G>T, p.Ser768Ile is the same amino-acid change previously reported in tumor specimens and clinical reports, and ClinVar variation 45251 records the exact variant as Pathogenic with criteria provided by a clinical laboratory. PMID:15623594 explicitly identifies S768I in EGFR exon 20, and PMID:23242437 reports S768I in clinical response contexts. SpliceAI predicts a maximum delta score of 0.022, so the available evidence does not indicate that the observation is primarily explained by altered splicing.

Rule

Same amino acid change as a previously established pathogenic variant regardless of nucleotide change. ... Caveat: Beware of changes that impact splicing rather than at the amino acid/protein level.

Evidence considered

• ClinVar variation 45251: NM_005228.5(EGFR):c.2303G>T (p.Ser768Ile), Pathogenic, criteria provided by a single clinical laboratory.
• PMID:15623594, Results, EGFR Mutation Spectrum: S768I was found in two tumors after direct sequencing of EGFR exons 18-21.
• PMID:23242437: EGFR-S768I alone was reported in a gefitinib-responsive NSCLC, and compound G719A+S768I was reported in two patients with partial responses to erlotinib.
• SpliceAI Lookup: max delta score 0.022 for NM_005228.5:c.2303G>T.

Sources

Reasoning

NM_005228.5:c.2303G>T (EGFR p.Ser768Ile) is absent from gnomAD v2.1 and gnomAD v4.1. The available non-cancer checks also report absence. This satisfies the generic PM2 population requirement of an allele frequency <= 0.0001, applied at supporting strength under the supplied ClinGen SVI threshold.

Rule

PM2 (supporting): <= 0.0001 — SVI recommendation to downgrade PM2 to Supporting (ClinGen SVI 2020) (PMID:25741868)

Evidence considered

• gnomAD v2.1 reports the variant as absent; this corresponds to an observed allele frequency of 0 in the queried dataset.
• gnomAD v4.1 reports the variant as absent; this corresponds to an observed allele frequency of 0 in the queried dataset.
• The available gnomAD v2.1 non-cancer and v3.1 non-cancer checks also report absence. The supplied generic PM2 calibration is <= 0.0001 at supporting strength (ClinGen SVI recommendation, PMID:25741868).

Sources

Reasoning

PP3 is applicable because NM_005228.5:c.2303G>T is a missense variant, EGFR p.Ser768Ile, so the governing generic path is REVEL only; SpliceAI and BayesDel are not used for this criterion. The REVEL score is 0.765, which meets the supplied generic ACMG/ClinGen-SVI PP3 supporting threshold of >= 0.644 (ClinGen SVI REVEL calibration, Pejaver et al. 2022, PMID:36413997).

Rule

PP3: Multiple lines of computational evidence support a deleterious effect on the gene or gene product (conservation, evolutionary, splicing impact, etc.). ... PP3 (supporting): >= 0.644 — ClinGen SVI REVEL calibration (Pejaver et al. 2022, Am J Hum Genet) (PMID:36413997)

Evidence considered

• Variant consequence is missense: NM_005228.5:c.2303G>T produces NP_005219.2:p.(Ser768Ile) in exon 20.
• Local REVEL lookup reports score 0.765 for chr7:55181312 G>T.
• Applied the supplied generic PP3 supporting cutoff REVEL >= 0.644, from ClinGen SVI REVEL calibration (Pejaver et al. 2022, PMID:36413997).
• SpliceAI max delta 0.022 was not counted because the governing rule selects the REVEL path for a missense variant; BayesDel was not used because no generic ACMG-strength calibration is established.

Sources

Reasoning

The governing PP4 rule requires a patient phenotype or family history highly specific for a disease with a single genetic etiology. The available reports describe non-small-cell lung cancer, including adenocarcinoma or adenosquamous carcinoma, but this is a heterogeneous cancer phenotype with multiple possible molecular causes and is not highly specific for this EGFR variant or a single genetic etiology.

Rule

Patient’s phenotype or family history is highly specific for a disease with a single genetic etiology.

Evidence considered

• Huang et al. 2004, abstract and Table 1: the reported cases had NSCLC, specifically adenocarcinoma or adenosquamous carcinoma; the paper does not establish a phenotype specific to a single genetic etiology.
• PMID:23242437, Discussion: the exact variant was reported in NSCLC patients, including compound EGFR mutations, without evidence that the clinical phenotype is specific to one genetic cause.

Sources

Reasoning

The governing PP5 definition concerns a reputable source reporting the variant as pathogenic when the evidence is unavailable for independent evaluation. The exact-variant ClinVar record is Pathogenic but has no expert-panel submission; it is supported by a single clinical laboratory submission and a separate submission without assertion criteria. Under the supplied generic PP5 restriction, an ordinary laboratory submission, aggregate label, or non-expert assertion cannot trigger PP5.

Rule

Reputable source recently reports variant as pathogenic, but the evidence is not available to the laboratory to perform an independent evaluation.

Evidence considered

• ClinVar Variation ID 45251, exact variant NM_005228.5(EGFR):c.2303G>T (p.Ser768Ile): Pathogenic, criteria provided, single submitter; expert-panel submissions = 0.
• ClinVar submission audit: SCV000062079 is a non-expert clinical-testing submission, and SCV002506984 has no assertion criteria and is not an expert-panel submission.
• The available ClinVar classification is an assertion that can be independently evaluated from the cited literature, and therefore does not satisfy the supplied generic PP5 restriction.

Sources

Reasoning

NM_005228.5:c.2303G>T (EGFR p.Ser768Ile) is absent from gnomAD v2.1 and gnomAD v4.1, including the available non-cancer population checks. Therefore no allele frequency or filtering allele frequency is present, and the generic BA1 threshold of >= 0.05 is not met.

Rule

BA1 (stand-alone): >= 0.05 — ACMG/AMP 2015 stand-alone benign allele frequency (PMID:25741868)

Evidence considered

• gnomAD v2.1 reports the variant as absent; no allele count, allele frequency, popmax AF, or grpmax FAF is available.
• gnomAD v4.1 reports the variant as absent; no allele count, allele frequency, popmax AF, or grpmax FAF is available.
• The available gnomAD v2.1 non-cancer and v3.1 non-cancer checks also report absence. The generic stand-alone BA1 calibration supplied for this case is >= 0.05 (ACMG/AMP 2015, PMID:25741868).

Sources

Reasoning

The variant is absent from gnomAD v2.1 and gnomAD v4.1, so no filtering allele frequency (grpmax FAF) is available that could exceed the generic BS1 default threshold of >= 0.01. There is consequently no population-frequency evidence that the allele is more common than expected for the disorder.

Rule

BS1 (strong): >= 0.01 — Generic SVI default; per-gene derivation preferred (Whiffin et al. 2017) (PMID:28518168)

Evidence considered

• gnomAD v2.1 reports absence of the variant and provides no allele count, allele frequency, popmax AF, or grpmax FAF.
• gnomAD v4.1 reports absence of the variant and provides no allele count, allele frequency, popmax AF, or grpmax FAF.
• The available gnomAD v2.1 non-cancer and v3.1 non-cancer checks also report absence. The generic BS1 default supplied for this case is >= 0.01; no disorder-specific maximum credible allele-frequency calculation is needed because no elevated frequency is observed.

Sources

Reasoning

The available gene and variant evidence does not support a gene in which disease is primarily caused by truncating variants: EGFR p.Ser768Ile is a documented missense alteration, the case materials describe EGFR missense gain-of-function/oncogenic biology, and the cited EGFR cancer literature reports recurrent kinase-domain missense variants. No supplied source shows that EGFR disease is primarily caused by truncating variants. Accordingly, the BP1 premise is not supported for this missense variant.

Rule

Missense variant in a gene for which primarily truncating variants are known to cause disease.

Evidence considered

• OncoKB describes EGFR p.S768I in a gain-of-function/oncogenic context.
• PMID:15623594 reports EGFR missense mutations in kinase-domain exons, including S768I.
• PMID:23242437 reports EGFR missense S768I-containing tumors and responses to EGFR inhibitors; these materials do not support a primarily truncating EGFR disease mechanism.

Sources

Reasoning

BP4 is applicable because NM_005228.5:c.2303G>T is a missense variant, but its governing generic path is REVEL only. The REVEL score is 0.765, which is not <= 0.29, the supplied generic BP4 supporting cutoff; therefore BP4 is not met. The SpliceAI max delta of 0.022 is not considered because a splice prediction cannot substitute for the applicable missense REVEL path, and BayesDel has no generic ACMG-strength calibration.

Rule

BP4: Multiple lines of computational evidence suggest no impact on gene or gene product (conservation, evolutionary, splicing impact, etc). ... BP4 (supporting): <= 0.29 — ClinGen SVI REVEL calibration (Pejaver et al. 2022, Am J Hum Genet) (PMID:36413997)

Evidence considered

• Variant consequence is missense: NM_005228.5:c.2303G>T produces NP_005219.2:p.(Ser768Ile) in exon 20.
• Local REVEL lookup reports score 0.765 for chr7:55181312 G>T.
• Applied the supplied generic BP4 supporting cutoff REVEL <= 0.29, from ClinGen SVI REVEL calibration (Pejaver et al. 2022, PMID:36413997); 0.765 does not meet it.
• SpliceAI max delta 0.022 was not counted because the governing rule selects the REVEL path for a missense variant; BayesDel was not used because no generic ACMG-strength calibration is established.

Sources

Reasoning

The governing BP6 definition concerns a reputable source reporting the variant as benign when the evidence is unavailable for independent evaluation. The exact-variant ClinVar record contains no expert-panel benign or likely-benign classification and instead contains a non-expert Pathogenic assertion plus a submission without assertion criteria. Under the supplied generic BP6 restriction, no BP6 evidence is present.

Rule

Reputable source recently reports variant as benign, but the evidence is not available to the laboratory to perform an independent evaluation.

Evidence considered

• ClinVar Variation ID 45251 is an exact-variant record with no expert-panel submissions; its available clinical-laboratory classification is Pathogenic, not benign or likely benign.
• ClinVar submission audit reports expert_panel_submissions = 0 and no benign or likely-benign expert-panel assertion.

Sources

Reasoning

No proband-specific de novo observation is documented. The available evidence does not establish that the variant is absent in both biological parents, does not confirm maternity and paternity, and does not document a disease-appropriate proband or absence of family history; therefore PS2 cannot be assessed.

Rule

De novo (both maternity and paternity confirmed) in a patient with the disease and no family history.

Evidence considered

• The case evidence contains no proband-specific parental testing result or documented maternity and paternity confirmation.
• The available case materials do not document the proband's disease phenotype or family history.

Still missing

• Proband-specific clinical diagnosis and phenotype are not documented.
• Maternal and paternal testing results are not documented.
• Family history is not documented.

Reasoning

The available variant-specific literature documents EGFR S768I in tumors, increased phosphorylated EGFR in one tumor, clinical responses to EGFR inhibitors, and a reported in-vitro result for the compound G719S+S768I construct. These observations are biologically relevant but do not establish that the exact single-nucleotide EGFR S768I substitution was tested in a validated, reproducible, robust functional assay with adequate controls. The evidence therefore does not support applying PS3 under the required well-established functional-study definition.

Rule

Well-established in vitro or in vivo functional studies supportive of a damaging effect on the gene or gene product.

Evidence considered

• Huang et al. 2004 (PMID:15623594), Results and Table 1, Cases 38 and 39: EGFR S768I was identified in two untreated NSCLC tumors by direct sequencing; increased phosphorylated EGFR was observed in Case 39 but not Case 38. This is tumor-observation evidence, not a validated controlled functional assay of the exact variant.
• Compound EGFR mutations and response to EGFR tyrosine kinase inhibitors (PMID:23242437), Discussion: the paper reports clinical responses for G719A+S768I and summarizes an in-vitro study of G719S+S768I as gefitinib-sensitive. The cited in-vitro result concerns a compound construct and is not sufficient here to establish a validated, robust assay for the exact single S768I substitution.

Still missing

• A directly tested exact EGFR S768I construct in a well-established functional assay with appropriate wild-type, pathogenic, and benign controls is not documented.
• Assay validation, reproducibility, quantitative effect size, and experimental controls sufficient for PS3 strength calibration are not documented.

Sources

Reasoning

The governing PS4 rule requires a significant increase in variant prevalence in affected individuals versus controls, with OR/RR >5.0 and a confidence interval excluding 1. PMID:15623594 verified the exact S768I variant in 2 of 101 untreated NSCLC tumors, but the study provides no unaffected or population control count for S768I. The cohort is a somatic tumor series, and gnomAD v2.1 and v4.1 report the variant as absent without a usable matched-ancestry denominator or case-control design. Therefore, a valid case-control OR and 95% CI could not be computed; the reported 2/101 case frequency alone cannot establish PS4.

Rule

The prevalence of the variant in affected individuals is significantly increased compared to the prevalence in controls. Note 1: Relative risk (RR) or odds ratio (OR), as obtained from case-control studies, is >5.0 and the confidence interval around the estimate of RR or OR does not include 1.0.

Evidence considered

• Huang et al. 2004, Results, EGFR Mutation Spectrum; Table 1, Cases 38 and 39: S768I was found in 2 tumors among 101 untreated NSCLC specimens; this is a case-only tumor series with no unaffected or population control comparison.
• The exact-variant paper was checked in its abstract and full text: the study reports 2/101 tumor cases carrying S768I, not a control count or a case-control estimate.
• gnomAD v2.1 and gnomAD v4.1 both report the exact variant as absent, but the available records do not provide a matched-ancestry control denominator suitable for a valid OR/RR and 95% CI calculation.
• No OR or 95% CI was computed because the required control count or frequency was not available.

Still missing

• A matched control cohort or matched-ancestry control allele frequency with usable denominators is required to calculate OR/RR and its 95% CI.

Sources

Reasoning

The variant lies at EGFR residue 768 in the kinase domain, and the case materials report that EGFR exons 18-21 encode the kinase domain. However, the generic PM1 rule also requires a mutational hotspot and/or critical well-established functional domain without benign variation. The materials do not provide a gene-level benign-variation assessment for the EGFR kinase domain. Cancer Hotspots and COSMIC occurrence cannot by themselves establish the generic PM1 requirement under the governing instructions because no VCEP hotspot route applies.

Rule

Located in a mutational hot spot and/or critical and well-established functional domain (e.g. active site of an enzyme) without benign variation.

Evidence considered

• PMID:15623594 describes EGFR exons 18-21 encoding the kinase domain and reports S768I in exon 20.
• Cancer Hotspots reports a statistically significant EGFR S768 hotspot, but the extracted result is partially parsed and no VCEP hotspot route applies.
• COSMIC records EGFR c.2303G>T (COSV51768106) 74 times in somatic cancers; this does not establish that the domain lacks benign variation.

Still missing

• A verified assessment of benign variation across the EGFR kinase domain is required to complete generic PM1.

Sources

Reasoning

The target is a missense change at EGFR Ser768, but the supplied evidence verifies S768I rather than a different missense substitution at Ser768 determined pathogenic. The PM5 candidate artifact contains no same-residue comparator candidates, and the extracted literature does not document a qualifying alternate Ser768 missense variant. Therefore comparator evidence is insufficient for a definitive PM5 decision.

Rule

Novel missense change at an amino acid residue where a different missense change determined to be pathogenic has been seen before. ... Caveat: Beware of changes that impact splicing rather than at the amino acid/protein level.

Evidence considered

• PM5 candidate collection: zero same-residue candidates and no verified alternate pathogenic missense comparator.
• PMID:15623594 reports S768I but not a different pathogenic missense substitution at Ser768.
• PMID:23242437 reports S768I and compound variants involving G719, but not a qualifying different pathogenic missense change at Ser768.

Still missing

• A verified pathogenic alternate missense variant at EGFR Ser768 is needed to assess PM5.

Sources

Reasoning

No assumed de novo observation is documented for this patient. There is no report of the variant in the proband with parental testing insufficient to confirm one or both biological relationships, so PM6 cannot be assessed.

Rule

Assumed de novo, but without confirmation of paternity and maternity.

Evidence considered

• The case evidence contains no proband-specific de novo or assumed-de-novo report and no parental testing details.

Still missing

• A proband-specific variant result is not documented.
• Parental testing or the specific missing maternity/paternity confirmation is not documented.
• The proband's disease phenotype is not documented.

Reasoning

No affected family members, informative meioses, or variant genotypes in relatives are documented. The available literature reports somatic tumor observations rather than familial segregation, so PP1 cannot be assessed.

Rule

Co-segregation with disease in multiple affected family members in a gene definitively known to cause the disease.

Evidence considered

• The available case materials contain no familial segregation dataset, affected-relative genotypes, or informative meioses.
• The extracted literature evidence concerns EGFR S768I in non-small-cell lung-cancer tumor specimens and does not provide familial co-segregation evidence.

Still missing

• No family pedigree or affected-relative phenotypes are documented.
• No relative genotypes are documented.
• No informative meioses or segregation count is documented.

Reasoning

The supplied materials establish that EGFR p.Ser768Ile is a recurrent somatic missense alteration with gain-of-function/oncogenic context, but they do not provide the gene-level benign missense rate or a systematic disease-variant spectrum needed to show both that EGFR has a low rate of benign missense variation and that missense variants are a common disease mechanism under generic ACMG/AMP.

Rule

Missense variant in a gene that has a low rate of benign missense variation and where missense variants are a common mechanism of disease.

Evidence considered

• OncoKB provides EGFR variant-specific gain-of-function and oncogenic context, but not the generic PP2 gene-level benign-missense threshold.
• PMID:15623594 and PMID:23242437 document recurrent EGFR missense alterations in NSCLC, but do not establish the required low benign missense rate.

Still missing

• A validated EGFR benign-missense constraint measure and systematic evidence that missense variants are a common disease mechanism are required.

Sources

Reasoning

The available evidence does not document the variant in a healthy adult individual, nor does it provide a validated healthy-adult cohort with genotype and age information sufficient to assess homozygous, heterozygous, or hemizygous observation. gnomAD absence is not evidence that BS2 is met or not met.

Rule

BS2

Still missing

• A reliable report of the variant in a healthy adult, with genotype, phenotype, age, and relevant inheritance context, is not available.
• No adequately characterized cohort is available to assess whether healthy heterozygous, homozygous, or hemizygous individuals were observed.

Reasoning

No variant-specific well-established in vitro or in vivo study in the available evidence demonstrates that EGFR S768I has no damaging effect on EGFR protein function or splicing. The available reports describe tumor occurrence, phosphorylation in one tumor, inhibitor response, or a compound-mutant assay, none of which establishes a benign functional result for the exact variant under appropriate controls.

Rule

Well-established in vitro or in vivo functional studies show no damaging effect on protein function or splicing.

Still missing

• A direct exact-variant assay demonstrating preserved EGFR protein function or normal splicing with appropriate controls is not documented.

Reasoning

No affected family members have documented negative testing for the variant, and no pedigree or non-segregation analysis is available. Therefore lack of segregation cannot be evaluated and BS4 cannot be assessed.

Rule

Lack of segregation in affected members of a family.

Evidence considered

• The case evidence contains no affected-relative genotypes or documented family-level non-segregation analysis.

Still missing

• No affected-relative testing results are documented.
• No pedigree, phenotype ascertainment, or segregation analysis is documented.
• Potential phenocopy and multiple-variant considerations cannot be evaluated.

Reasoning

BP2 cannot be assessed from the available evidence. The governing rule requires either observation in trans with a pathogenic variant in a fully penetrant dominant gene/disorder or observation in cis with a pathogenic variant in any inheritance pattern. PMID:15623594 reports S768I together with G719C or G719S in two tumors, and PMID:23242437 reports compound G719A+S768I, but neither source establishes cis or trans phase, and the provided evidence does not independently establish the accompanying EGFR variant as pathogenic for this rule. The observations are somatic tumor findings rather than affected-proband segregation or inherited phase data.

Rule

Observed in trans with a pathogenic variant for a fully penetrant dominant gene/disorder or observed in cis with a pathogenic variant in any inheritance pattern.

Evidence considered

• PMID:15623594, Results and Table 1, reports S768I in Cases 38 and 39 together with G719C or G719S, respectively, in directly sequenced NSCLC tumors; the extract does not report cis/trans phase.
• PMID:23242437, Discussion, reports compound EGFR-G719A+S768I in two NSCLC patients and summarizes additional compound-mutant reports; the extract does not establish molecular phase or independently classify G719A as pathogenic for BP2.

Still missing

• Direct phase evidence showing whether the accompanying EGFR variant is in cis or trans with S768I.
• Evidence that the accompanying EGFR variant is pathogenic under the applicable disorder context.
• Confirmation of a fully penetrant dominant inherited disorder if the trans pathway is being considered.

Sources

Reasoning

The governing BP5 rule requires the variant to be found in a case with an alternate molecular basis for disease. The reviewed reports document EGFR S768I in NSCLC tumors, including tumors with additional EGFR mutations, but do not establish an alternate molecular basis for the disease in the same case. An additional EGFR alteration is not by itself an established alternate molecular basis under the generic rule, and the available evidence does not provide sufficient case-level molecular workup to adjudicate BP5.

Rule

Variant found in a case with an alternate molecular basis for disease.

Evidence considered

• Huang et al. 2004, Table 1, Cases 38 and 39: S768I occurred with G719C or G719S in the same EGFR gene; the paper does not identify an alternate molecular basis for NSCLC.
• PMID:23242437: compound EGFR-G719A+S768I was reported in two NSCLC patients, but the publication does not establish that another molecular cause, independent of the EGFR findings, caused the disease.

Still missing

• Case-level evidence of a confirmed alternate molecular cause of the disease is required.

Sources